Author

Rob Hickey

Chief AI Officer · LexData · 64 articles

Transmission tower over a country road from a drone, six insulator strings and one corrosion mark boxed, from a customer inspection run

Computer vision · 6 min read

How to evaluate object detection models fairly, and why two mAP numbers are rarely comparable

A vendor's score and yours are answers to different questions. The test set, the calculation and the inference conditions each move the number on their own.

Rob Hickey · Oct 4, 2026

Warehouse aisle with racked pallets from a high camera, forklift and pallet boxed, generated scene with detections from our model

Computer vision · 6 min read

Overfitting in computer vision, or how a model learns the warehouse instead of the pallet

A pallet detector that shines at its own warehouse and fails at the next one learned the lighting, and the training versus validation gap is the tell.

Rob Hickey · Oct 4, 2026

Loading dock from a mounted camera, trucks at the bays and a forklift with a pallet, forklift, pallet, person and truck boxed, generated scene with detections from our model

Computer vision · 6 min read

How to train, evaluate and deploy a computer vision model on a camera you already own

From the first labeled frame off a dock camera to a model watching that dock, in the order the work really happens, with the retrain built in from day one.

Rob Hickey · Oct 4, 2026

Warehouse aisle from a high camera, a forklift and a racked pallet boxed, generated scene with detections from our model

Computer vision · 6 min read

What CLIP is and why typing a description finds the frame

A shared space for pictures and words lets an operator type a sentence and get the matching frames from a recorder. A way to look, and the boxes come after.

Rob Hickey · Oct 3, 2026

Crane lift over a fenced site with workers in hi-vis, hard hats and the load boxed, generated scene with detections from our model

Labeling · 6 min read

A dataset quality audit for object detection when mAP hides a weak class

The PPE camera at the rig site scored well on average and missed most bare heads. Audit the labels per class, fix the boxes in review, and retrain on the fixes.

Rob Hickey · Oct 3, 2026

Forward camera at a junction, vehicles, pedestrians and signs boxed, from a customer perception run

Computer vision · 7 min read

Monocular vs stereo depth estimation for a robot that has to stop in time

A stop rule needs a distance in metres. Monocular depth gives a ranking, stereo gives metres and loses them on a blank wall. The rule decides which one.

Rob Hickey · Oct 3, 2026

Robot arm loading a machine behind a safety fence, the fence and the arm boxed, generated scene with detections from our model

Computer vision · 7 min read

Pose estimation algorithms and what keypoints give a safety camera that boxes do not

A box says a person is beside the press. Keypoints say where the wrist is. How joints are learned as heatmaps, and which questions are worth the cost of asking.

Rob Hickey · Oct 3, 2026

Scaffold with two workers, guardrails, ladder and plank decks boxed, from a customer site camera

Computer vision · 6 min read

Transfer learning from a pretrained checkpoint, why a few hundred frames are enough when you start warm

A pretrained backbone already knows edges and shapes. A few hundred reviewed frames from a site camera teach the head what a guardrail is. Retrains begin warm.

Rob Hickey · Oct 3, 2026

Three transmission towers across a field, insulators boxed, defects flagged and vegetation marked near the line, from a customer aerial survey

Computer vision · 7 min read

What a vision transformer does with a frame, and what changes in monitoring when you use one

The frame becomes patches, every patch looks at every other, and a tower over a hedge gets easier. The cost grows with resolution. The review queue stays.

Rob Hickey · Oct 3, 2026

Stamping line with steel panels on a conveyor passing an inspection station, generated scene with detections from our model

Labeling · 6 min read

Blur augmentation in computer vision, training for the blur the camera will produce

A camera on a vibrating gantry, an autofocus that hunts, fog on the housing. Train on the blur the line makes, never on the class it would erase.

Rob Hickey · Oct 2, 2026

Food processing line, fillets on the conveyor, generated scene with detections from our model

Operations · 7 min read

Computer vision event logging that keeps the frame with the prediction

A missed bone fragment on the night shift can only be explained if the frame, the prediction, the model version and the lot were logged together.

Rob Hickey · Oct 2, 2026

Assembly station, housing and cable boxed, generated scene with detections from our model

Labeling · 7 min read

Handling imbalanced classes when the defect is one part in a thousand

A pinched cable once in a thousand housings. Why the loss can afford to ignore it, how to collect the rare frames on purpose, and the one honest metric.

Rob Hickey · Oct 2, 2026

Distribution poles along a desert track, one insulator flagged as cracked, from a customer inspection run

Labeling · 7 min read

Polygon image annotation or a box, decided by the question

Polygon image annotation, a box or a mask. A cracked insulator, a corroded patch and a person in a restricted zone each force a different shape.

Rob Hickey · Oct 2, 2026

Robotic welding cell, part on the fixture, generated scene with detections from our model

Operations · 6 min read

How big a computer vision model the line camera actually needs

Nano to large on one weld cell camera. What a bigger model buys on look-alike defects, what it costs on the runner, and choosing by the device it runs on.

Rob Hickey · Oct 2, 2026

Factory floor with the robot cell fence boxed, generated scene with detections from our model, the kind of frame a crew inside the fence turns unfamiliar

Operations · 7 min read

Embedding based anomaly detection for the frame the model has never seen

A car park camera meets its first snowplough and a line camera meets the maintenance crew. Distance from the usual frames is one check that finds them.

Rob Hickey · Oct 1, 2026

Edge runner beside a recorder, generated scene with detections from our model

Operations · 6 min read

How to monitor inference health when nothing throws an error

A camera that stops sending, a runner that falls behind, a correction rate climbing on one site. Three failures with no error message and three different fixes.

Rob Hickey · Oct 1, 2026

Loading dock with a truck backed in, generated scene with detections from our model

Operations · 7 min read

Motion detection in computer vision that ignores the parked truck

A dock camera that alerts on the truck that just arrived and stays silent on the one parked since 5 am, with wind in the trees trained out.

Rob Hickey · Oct 1, 2026

Shopper at a dairy case with the basket and rack sections boxed, from a customer store camera

Operations · 6 min read

How to reduce class flickering in object detection on video

A shelf facing that reads empty, stocked, empty every few frames is a per-frame model at a boundary. Vote over the track, then review the frames that flickered.

Rob Hickey · Oct 1, 2026

A loading dock from a mounted camera, trucks at the bays and a forklift with a pallet boxed, generated scene with detections from our model

Operations · 6 min read

YOLO versus vision language models, decided by the failure you can afford

A detector gives the same box for the same frame, every two seconds, forever. A language model answers a question nobody wrote a class for. The dock needs both.

Rob Hickey · Oct 1, 2026

Assembly station, housing and fasteners laid out with a cable boxed, generated scene with detections from our model

Industries · 7 min read

Diabetes device inspection with a VLM and one detector across four device shapes

Four device shapes on one line: a question judges the faults nobody wrote a class for, a detector owns the count, and a new device generation is labeled first.

Rob Hickey · Sep 30, 2026

Two workers under netting on a scaffold, hard hats boxed, from a customer site camera

Operations · 6 min read

Robustness testing for computer vision ahead of the weather

Robustness testing for computer vision means you blur, darken, shift and compress the held-out frames before the October rain does, then read recall.

Rob Hickey · Sep 30, 2026

Dashcam frame at a junction, vehicles, pedestrians and signs boxed, from a customer perception run

Industries · 6 min read

Self-driving car perception and the white trailer against a bright sky

A model cannot find what its labels never showed it, the miss looks like nothing on the frame, and doubted fleet frames are how the training set catches up.

Rob Hickey · Sep 30, 2026

A robotic welding cell with a part on the fixture, generated scene with detections from our model

Operations · 6 min read

Underspecification in machine learning, and why two models with the same score do not both survive the second plant

Fifty training runs, one recipe, one validation score. The test set could not tell them apart, and the second plant's cameras can.

Rob Hickey · Sep 30, 2026

Transmission line over farmland, encroaching vegetation marked near the conductors, from a customer aerial survey

Industries · 7 min read

Wildfire smoke detection from a fixed ridge camera at dawn

Smoke is a class with no edges, cloud and valley mist are the false alarms to label, and a camera that never moves is what makes the first thin column visible.

Rob Hickey · Sep 30, 2026

Robotic welding cell, a part on the fixture and the robot arm boxed, generated scene with detections from our model

Industries · 7 min read

Zero-shot pose estimation for a collaborative robot, and what makes it hold

A generalist pose model reads a worker's reach on day one, keypoints are smoothed across sampled frames, and the site's own reviewed frames make it hold.

Rob Hickey · Sep 30, 2026

Transmission line over farmland, towers, insulators and encroaching vegetation boxed, from a customer aerial survey

Industries · 7 min read

Aerial fire detection from a drone patrol, smoke before the flame reaches the line

On a right-of-way patrol, smoke is a few dozen grey pixels that look like haze. Two boxed classes, an alert with a cooldown, and the bad-weather days kept.

Rob Hickey · Sep 29, 2026

Packhouse sorting conveyor, apples boxed by the detector, generated scene

Industries · 6 min read

Automated sorting with computer vision, from the camera over the conveyor to the diverter

A box on every apple, a grade from the box, and an air jet that acts on it before the belt moves on. The new cultivar is when the model needs the graders again.

Rob Hickey · Sep 29, 2026

Street from a dashcam, traffic signs, pedestrians and vehicles boxed, from a customer perception run

Industries · 6 min read

Infrastructure asset management with computer vision from a truck at road speed

Signs, guardrails and poles pass the truck camera in a fraction of a second. The model finds them, the register says what should be there, a crew gets the gap.

Rob Hickey · Sep 29, 2026

Packaging line with cartons and labels boxed as they pass the scanner, generated scene with detections from our model

Operations · 7 min read

Benchmark the model on your own cameras before you believe a score

Two candidate models, two published scores, and a packaging line that only cares which one finds the torn label. The held-out week from your cameras decides.

Rob Hickey · Sep 28, 2026

Stamping line with steel panels passing an inspection station, press and conveyor boxed, generated scene with detections from our model

Computer vision · 6 min read

Reading a training run before trusting the model

The loss curve says the run finished. The gap to validation, the recall on the crack class and the frames behind the score say whether the line can trust it.

Rob Hickey · Sep 28, 2026

Dairy case with three empty slots flagged across the rack sections, from a customer store camera

Operations · 6 min read

The class that fails behind a good average, read off the confusion matrix

A shelf model with a healthy mAP and a dairy case full of unreported gaps. The matrix names the cell the failing class fell into, and the cell names the fix.

Rob Hickey · Sep 28, 2026

Milk shelf with three empty slots flagged and the rack sections boxed, from a customer store camera

Operations · 7 min read

Precision, recall and the mistakes each one lets through

A shelf model that cries wolf and an inspection model that waves the crack past are wrong in opposite ways. The threshold is the dial between them.

Rob Hickey · Sep 28, 2026

Assembly station with a housing, fasteners and a cable loom laid out, generated scene with detections from our model

Labeling · 6 min read

Out-of-scope detection, what the model does when it sees something it was never taught

A screw detector on the phone line meets an operator's hand and calls it a screw. Null frames, a class for confusing things, a test set from the real camera.

Rob Hickey · Sep 28, 2026

Camera on a pole over an industrial yard at dusk, camera, fence and vehicle boxed, generated scene with detections from our model

Computer vision · 6 min read

Real-time object detection models, transformer or convolutional, and which size for the runner

A DETR-style detector drops the anchors and the suppression step. The size to run is decided by the box beside the camera, which the leaderboard cannot see.

Rob Hickey · Sep 28, 2026

Stamping line, steel panels on the conveyor passing the inspection station, generated scene with detections from our model

Labeling · 7 min read

What active learning is, and why it labels the frames that change the model

Fifty thousand frames of good panels and nine splits. Active learning sends the frames the model doubts to a person first, and the cycle runs on the live line.

Rob Hickey · Sep 28, 2026

Robotic welding cell from the cell camera, part and robot arm boxed, generated scene with detections from our model

Operations · 7 min read

Active learning for computer vision on a line camera that never stops

The weld camera runs three shifts. The model returns the frames it doubts, the inspector corrects them, and past the threshold a new version trains and ships.

Rob Hickey · Sep 27, 2026

Packaging line past the scanner, cartons and conveyor boxed, generated scene with detections from our model

Operations · 7 min read

How to improve computer vision model accuracy by starting with one defect class

A coffee bag line wanted eight defect classes and got none. Torn bag first, frames from the station camera, a written guideline, then the loop adds the rest.

Rob Hickey · Sep 27, 2026

Food processing line with fillets boxed on the blue conveyor, generated scene with detections from our model

Operations · 6 min read

Human-in-the-loop computer vision, the mistakes the model is sure about and the inspector who catches them

A quality inspector flags a wrong detection in seconds, the confidently wrong frames no threshold would catch, and the correction rate as the signal.

Rob Hickey · Sep 27, 2026

Bread shelf with an empty slot flagged and the rack sections boxed, from a customer store camera

Operations · 6 min read

Monitoring a computer vision model in production, and what to do when the corrections start climbing

A shelf model meets the Christmas displays: the correction rate as the signal, doubted frames first in the queue, a new version compared on the same frames.

Rob Hickey · Sep 27, 2026

Mast camera over a container yard at dusk, generated scene with detections from our model

Operations · 8 min read

Monitoring data drift in computer vision, and the one signal that catches both kinds

A container yard camera at night is data drift. A damage policy that now counts dents is concept drift. Only the correction rate catches both.

Rob Hickey · Sep 27, 2026

Rack scan down a clothing aisle after a seasonal reset, garment stacks boxed, from a customer store camera

Operations · 7 min read

Production deployment is the halfway point of a computer vision model's life

A packaging line adds a flavour and a store resets for autumn. The model that scored well in evaluation meets both on a Monday, and the second half begins.

Rob Hickey · Sep 27, 2026

Orchard tree rows and drivable paths, from a customer tractor camera

Industries · 6 min read

AI crop analysis in the greenhouse, catching tomato disease before it spreads down the row

Lesions boxed by disease with a healthy class, a question of the footage about how far a patch spread since last week, and a model that turns with the season.

Rob Hickey · Sep 26, 2026

Cartons with labels on a packaging line, carton and label boxed, generated scene with detections from our model

Industries · 6 min read

Juice box quality inspection at the straw applicator, with a tag on every carton

Acceptable, missing, bent or pierced on every carton from the line's own frames, and why a nudged camera mount is the failure a classifier shows first.

Rob Hickey · Sep 26, 2026

Conveyor and press on a stamping line, generated scene with detections from our model

Industries · 7 min read

Predictive maintenance with computer vision on the bearing that is about to seize

A camera on the bearing housing, the belt and the pipe joint sees the weep, the fray and the loose bolt weeks before the stop, and sends the frame.

Rob Hickey · Sep 26, 2026

A robot arm loading a machine behind a safety fence, generated scene with detections from our model

Industries · 7 min read

Robotic pick and pack with a bounding box and a mask on every item, and why the arm still misses

A box finds the item, a mask finds where the cup should land, foil packaging punches holes in the depth map, and a nudged camera mount shows up as failed picks.

Rob Hickey · Sep 26, 2026

Steel sheet on a stamping line conveyor passing an inspection station, generated scene with detections from our model

Industries · 8 min read

Computer vision in steel manufacturing, from the ladle shell to the coil

A thermal camera on the ladle stand sees the lining fail from the outside, and a replaced camera is the day the hotspot map has to be redrawn.

Rob Hickey · Sep 25, 2026

Bottling line with bottles queued under a fill head, generated scene with detections from our model

Industries · 7 min read

Contact lens defect inspection with pass, review and fail as three answers

A dark-field camera turns a crack in a clear lens into a bright line. The lens the model cannot decide goes to a person, and that middle bucket is the product.

Rob Hickey · Sep 25, 2026

Wind turbine blade from a drone, a chip in the leading edge, generated scene with detections from our model

Industries · 7 min read

How to detect small defects with computer vision when three resizes erase them

A hairline crack on a blade and a micro-scratch on a display vanish before the first layer sees them. Tiling keeps their pixels; the labeler sees them first.

Rob Hickey · Sep 25, 2026

Lattice tower from a drone pass with insulator strings boxed and corrosion flagged on the hardware, from a customer inspection run

Industries · 6 min read

Drone structural damage detection with cracks and spalling confirmed by the engineer

Masks for cracks and spalling from a close drone pass, so area compares pass to pass. The engineer confirms each flag, and a re-planned flight moves every mask.

Rob Hickey · Sep 25, 2026

Loading dock with a forklift, a pallet and a person boxed, generated scene with detections from our model

Industries · 7 min read

Fall detection with pose estimation on the warehouse cameras a plant already has

A worker down behind the racking at 2 am, a torso line that drops between samples, and an alert with the frame attached that reaches someone who can walk there.

Rob Hickey · Sep 25, 2026

Robotic welding cell, generated scene with detections from our model

Industries · 6 min read

Flaw detection with computer vision for the flaws an inspector stops seeing at hour eight

Surface, structural and material flaws on one welded bracket, why inspectors disagree by hour eight, and the review loop that holds the 7 am standard.

Rob Hickey · Sep 25, 2026

Welding cell, a part on the fixture with the robot arm boxed, generated scene with detections from our model

Industries · 6 min read

AI visual inspection as the nondestructive testing step a camera can take over

Visual testing is the first NDT gate, its acceptance criteria are already written, and a camera can apply them to every weld instead of one in twenty.

Rob Hickey · Sep 24, 2026

Packaging line past a scanner, generated scene with detections from our model

Industries · 6 min read

Computer vision pill inspection catches one damaged tablet in a thousand

A camera over the tablet conveyor before the blister packer, recall set on the defect side, and thresholds re-tuned when a new batch moves the base rate.

Rob Hickey · Sep 24, 2026

Stamping line, panels passing an inspection station, generated scene with detections from our model

Industries · 6 min read

Machine vision or computer vision on the line, and when each breaks

A threshold rule that died with the old fluorescent tubes, a template that broke on the new variant, and the model that is corrected rather than re-engineered.

Rob Hickey · Sep 24, 2026

Industries · 6 min read

On-shelf availability monitoring with the aisle camera a store already has

Empty facings boxed on the bread bay, glare and black packaging as the false alarms, a cooldown rule to the restocker, alert volume watched against itself.

Rob Hickey · Sep 24, 2026

Robotic weld cell at the torch, generated scene with detections from our model

Labeling · 6 min read

Writing a class ontology for defect labeling before the first box

On a weld line, the boundary, occlusion and inclusion rules come before the first box. When the tolerance tightens, the old labels are wrong.

Rob Hickey · Sep 23, 2026

Analog pressure gauge on a plant pipe, generated scene with detections from our model

Industries · 8 min read

Reading analog gauges with a camera, without replacing them

A camera on the dial turns a clipboard round into a stream of readings, and a nudged bracket is what quietly breaks it.

Rob Hickey · Sep 23, 2026

Robotic welding cell with the torch and the robot boxed over the fixture, generated scene with detections from our model

Industries · 6 min read

Reducing manufacturing scrap with a camera on the line, because scrap is a timing problem

A torch drifts at 9 am and the shift finds out at 2 pm. Scrap scales with that lag, and the camera flags the first bad bead rather than the hundredth.

Rob Hickey · Sep 23, 2026

Industries · 6 min read

Red zone monitoring with a camera over the forklift lane

Person and vehicle classes trained on the site's own frames, the zone as a polygon, dwell as the rule, and a nudged camera as what silently moves the zone.

Rob Hickey · Sep 23, 2026

Transmission line over farmland, insulators, defects and encroaching vegetation boxed, from a customer aerial survey

Industries · 7 min read

Transmission line inspection with object detection from a survey flight

Damaged strands, broken insulators, foreign objects and vegetation from one flight, tiled so a strand keeps its pixels, with recall set per hazard.

Rob Hickey · Sep 23, 2026

Food processing line with fillets on the blue conveyor and the rail boxed, generated scene with detections from our model

Industries · 6 min read

Visual anomaly detection on the line, catching the defect you never labeled

Learn normal from clean frames, score every fillet against it, route the strange ones to a person, and turn what they name into classes over time.

Rob Hickey · Sep 23, 2026

Bottling line, bottles queued under the fill head with one missing its cap boxed, generated scene with detections from our model

Computer vision · 7 min read

How to catch model drift before your customers do

A cap-check model passed capless bottles for six weeks and nobody knew until a retailer rang. The signal was there the whole time, in the review queue.

Rob Hickey · Jul 14, 2026

Lattice tower from a drone pass, insulator strings boxed and two corrosion spots flagged on the hardware, from a customer inspection run

Case notes · 7 min read

Corrosion detection on transmission insulators, from the inspection pass to a pin you can name

A drone pass along an extra-high-voltage line shows every shed and pin. The frame with early rust looks like the one before it; tracking makes the call stick.

Rob Hickey · Jul 2, 2026

Thermal camera on a fence line at night, two people at the fence, vehicles and the fence line boxed, from a customer site camera

Edge · 6 min read

Why your best vision model should live at the edge, next to the camera

A substation at the end of a long feeder has a thermal camera, a cellular uplink that drops in rain, and a model that keeps watching when the link does not.

Rob Hickey · May 22, 2026